PTSD is a chronic and distressing disorder, affecting up to 8 percent of the U.S. population. It is associated with significant morbidity, possible increased mortality and may persist for many years. Controlled studies have shown efficacy for antidepressant drugs in PTSD. Specifically, fluoxetine has been the most studied. No trials have evaluated the effects of long-term pharmacotherapy in PTSD, nor have they looked at the issues of relapse upon discontinuation. Little is known as to the optimum length of pharmacotherapy, or even whether long-term therapy is necessary. In this proposal the investigator aim to assess relapse and recurrence rates of PTSD in 74 patients with PTSD who have responded successfully to 6 months of open-label treatment with fluoxetine (FLU). One-hundred-and-twenty-six subjects will enter open-label treatment with FLU for 6 months. At 6 months, 82 subjects will be randomized to either continue on FLU or placebo (PBO) double-blind for a period of up to 6 months, at which time FLU responders will again be randomized to continue on FLU or receive PBO for a further 3 months, thus providing a total term of 15 months treatment in up to a maximum of 74 patients. The goals of the study are to compare rates of relapse/recurrence in the FLU and PBO groups after each of the discontinuation points. Subjects will initially be diagnosed by the SCID, with symptom severity measures including the CGI, structured interview (SI), Hamilton Scales and self-rating measures for PTSD, to include the Davidson Trauma Scale, Impact of Events Scale Revised and patient Global Improvement. Adverse effects of the medication will also be monitored and compliance checked with pill counts, daily logs and medication levels. The study will provide information about whether there is extended benefit by continuing pharmacotherapy for more than 6 months, whether long-term pharmacotherapy reduces the likelihood of symptom return during 12 months of pharmacotherapy. Health service and cost implications of such findings remain considerable, given the morbidity associated with PTSD.